Automating bus stop dwell time measurements for London buses using iBus

Abstract

iBus is Transport for London (TfL)’s GPS-enhanced Automatic Vehicle Location System, which has been rolled out to the entire contracted fleet of over 8,500 buses across London, and resulted in efficiencies in fleet management, improvements in bus waiting times, and provided improved real-time information for passengers. The System resides on board each vehicle, as well as in operators’ bus garages and at the main TfL Control Centre, and records a number of on-street events relating to for example buses’ entry and exit into stop zones, when their doors opened and closed, and their location and speed in real-time. This information, which is collected in the ‘log’ files of every vehicle, provided an opportunity to develop further uses for the System, including an alternative method for measuring bus stop dwell times. Historically, dwell times in London have been obtained using manual road-side surveys, which are relatively expensive, and therefore occur infrequently. However, dwell times and their variability are important to bus operations, network planning and traffic management, and they can affect the ability of urban traffic control systems such as SCOOT to provide buses with priority at traffic signals, which reduces their effectiveness. An alternative method for measuring dwell times using iBus therefore offers many benefits for TfL, provided a process could be determined and largely automated, as the dwell values are not recorded directly by the System, which is relatively complex. A knowledge base of the bus log files therefore had to be developed, and was tested to allow different algorithms, flow charts and programs to be produced for deriving dwell times, based on a sequence of different vehicle speed, stop zone and door events. An experiment was also conducted to validate the dwell times obtained through this method against video data obtained of vehicles stopping on street, which showed a close match between when the vehicle speeds are zero and roadside dwell, although another method, using the duration between when doors opened and closed, provided a close approximation, particularly when an offset value is accounted for. The dwell times obtained through the ‘speed zero’ method in the experiment were then analysed, and this showed wide variations between different bus stops and routes, which are consistent with previous surveys in London, and suggests that generalised values of dwell are inadequate for most applications. The analysis also showed that the dwell time variation by time-of-day is more complex than a traditional morning and afternoon peak, which may reflect changes in ticketing, vehicle modernisation, and the demand made by different types of bus passengers in recent years.

More information

Identifiers

Catalogue record

Export record